4 IT. S. BUREAU OF FISHERIES 



especially in the Panama region, where many geminate species are 

 found on each side of the Isthmus, which has split its species from 

 their mates since Miocene times. The Miocene fossil fishes of Cali- 

 fornia are largely geminate in origin, twins or parents of species now 

 appearing on the coast. In most cases such forms are clearly distinct 

 but to all appearance ancestral, the existing distinctions apparently 

 due to varying incidents of selection in varying environment. Gem- 

 inate species rarely inhabit the same region, though exceptions are 

 known, attributable to reinvasion. 



In many cases the distinctions between geminate forms are not 

 complete, intergradations occurring along the line of the barriers or 

 at different sections in climatic range. Types thus connected are 

 known as subspecies. Subspecies are often nascent species; but in 

 most cases a large amount of material is necessary before subspecies 

 can be separated from geminate species. Among fresh-water fishes, 

 each river basin shows some features of distinction, large or small, as 

 does each oceanic island with regard to marine fishes. In most 

 families of American river fishes studies of species have not been 

 extensive enough to establish facts of intergradation. While in 

 birds, reptiles, and mammals the subspecies are usually well defined 

 and named, among the fishes this is still largely a matter of guess- 

 work. Many authors base the recognition of subspecies not on 

 actual intergradation but on guesses based on degrees of probability, 

 a practice we have not followed. It has seemed to us better to regard 

 related forms as distinct species until intergradation has been shown. 



We find also among river fishes, and sometimes among marine 

 forms (as in Lutianus), occasional individuals that to all appearance 

 are hybrids. These are marked by a distinct blending of specific 

 traits. They are extremely rare, however, often known only from a 

 single example. A few hybrids are known or suspected among 

 certain minnows, sunfishes, and darters. The best-known hybrids 

 thus far recorded in America occur among the Centrarchidae, or 

 sunfishes, in which group they have been studied critically by Mr. 

 Hubbs. Two common species (Apomotis cyanellus and Eupomotis 

 gibbosus) figure particularly as probable parents. Hybrids may give 

 rise to distinct species, but no example of this is known among our 

 fishes. 



Among fishes there are also occasional monstrosities, freaks, or 

 "mutations," in which certain traits may be absent or exaggerated. 

 Whether such qualities may be inherited is not known. If so', it is 

 likely that they would soon be swamped by breeding with the mass 

 unless especially protected by segregation. It does not appear that 

 any species of fishes have arisen by mutation or from hybridization, 

 though the latter possibility is greater than the former. 



We may observe that compilations such as the present paper should 

 constitute an exercise in modesty. Such work can never be made 

 absolutely complete nor free from errors. It deals with the obser- 

 vations and thoughts of great minds, as also with a degree of ignor- 

 rance, carelessness, and perversity, by which names have been 

 multiplied without corresponding extension of ideas. Synonymy 

 (the use of different names for the same group) in general represents 

 the lack of exact knowledge or else a failure (often unavoidable) to 

 deal justly with the work of others. It is, therefore, as has been 

 said by Doctor Coues, often "a burden and disgrace to science." 



